international journal of sport psychology, 45 · pdf file1 international journal of sport...

International Journal of Sport Psychology, 45, 121-137. 1

doi: 10.7352/IJSP 2014.45.121 2

3

4

5

6

7

Collective efficacy or team outcome confidence? Development and validation of the 8

Observational Collective Efficacy Scale for Sports (OCESS) 9

10

11

Fransen, Katrien1, Kleinert, Jens

2, Dithurbide, Lori

3, Vanbeselaere, Norbert

4, & Boen, Filip

1. 12

13

1 Department of Kinesiology, KU Leuven, Leuven, Belgium 14

2 Institute of Psychology, German Sport University Cologne, Köln, Germany 15

3 School of Health and Human Performance, Dalhousie University, Halifax, Canada 16

4 Center for Social and Cultural Psychology, KU Leuven, Leuven, Belgium 17

18

E-mail address of lead author: [email protected] 19

20

21

22

Acknowledgments 23

This research was supported by a PhD Fellowship (Aspirant) of the Research Foundation 24

Flanders (FWO), awarded to Katrien Fransen. 25

http://www.kuleuven.be/wieiswie/en/unit/50000355

mailto:[email protected]

DEVELOPMENT AND VALIDATION OF THE OCESS 2

Abstract 26

Although collective efficacy has been demonstrated to be an important precursor of team 27

performance, there remains some ambiguity concerning its assessment. Therefore, the main 28

aim of the present study was to test the validity of previous collective efficacy measures. An 29

online survey was completed by 4,451 Flemish players and coaches from nine different team 30

sports. The results revealed two distinct and reliable scales; process-oriented collective 31

efficacy (i.e., the confidence in the team‟s skills to accomplish processes that could lead to 32

successes) and outcome-oriented team confidence (i.e., the confidence in the team‟s ability to 33

obtain a goal or win a game). Furthermore, we established the validity of a 5-item 34

Observational Collective Efficacy Scale for Sports (OCESS) as short measure of process-35

oriented collective efficacy. Because the OCESS only includes observable behaviors, this 36

scale has the potential to be a starting point for the development of a continuous in-game 37

measure of collective efficacy. 38

Keywords: instrument development, team confidence, continuous measure, team 39

sports, dynamic measurements, in-game variation 40


Collective efficacy or team outcome confidence? Development and validation of the 41

Observational Collective Efficacy Scale for Sports (OCESS) 42

The performance of athletes can vary strongly during a sports game. Players‟ 43

confidence in the team‟s capabilities is often mentioned as one of the factors that characterize 44

these performance variations throughout the game. For example, a sudden collapse in team 45

performance is often attributed to a drop in the team‟s confidence. Conversely, team 46

confidence is assumed to be a prerequisite for fighting back when the team is lagging behind. 47

Arsenal coach Arsene Wenger adds that “confidence is the easiest thing to lose in football and 48

the most difficult to win back” (Mangan, 2013). Bandura (1997, p. 477) termed this 49

confidence „collective efficacy‟, defined as “the group‟s shared belief in its conjoint capability 50

to organize and execute the courses of action required to produce given levels of attainment.” 51

Collective Efficacy as a Dynamic Construct 52

Bandura (1997) stated that collective efficacy has an effect on what a team chooses to 53

do, how much effort is instilled into a task, and how persistent the team is. These claims have 54

been supported in research showing that teams with strong collective efficacy beliefs tend to 55

set more challenging goals (Silver & Bufanio, 1996), exert more effort, and persist longer in 56

the face of adversity (Greenlees, Graydon, & Maynard, 1999). As a result, a positive 57

relationship has been revealed between collective efficacy and sport performance; the more 58

the players believe in the team‟s capacities, the better they perform and vice versa 59

(Dithurbide, Sullivan, & Chow, 2009; Hodges & Carron, 1992; Keshtan, Ramzaninezhad, 60

Kordshooli, & Panahi, 2010; Myers, Feltz, & Short, 2004; Myers, Payment, & Feltz, 2004). 61

It is important to emphasize that collective efficacy is not a fixed trait, but a dynamic 62

construct (Myers & Feltz, 2007). In other words, the individual‟s beliefs in the capabilities of 63

his or her team may change in the course of weeks, days, or even during a game. Especially 64

these changes in the course of a competition seem often responsible for winning or losing. To 65


investigate this close link between collective efficacy and performance, Bandura (1997, p. 67) 66

stated that “the relationship between efficacy beliefs and action is revealed most accurately 67

when they are measured in close temporal proximity.” Myers and colleagues (2007) added 68

that only research designs allowing for simultaneous measures of both efficacy and 69

performance would provide maximal information about their dynamic relationship during a 70

competition. However, in contrast with these guidelines and collective efficacy‟s dynamic 71

nature, the concept has traditionally been measured as a trait concept or at best before or after 72

a game, but not during a game. The only exception is a study by Edmonds, Tenenbaum, 73

Kamata, and Johnson (2009), who attempted to measure collective efficacy beliefs of 74

adventure racing teams at three time points during the race. Their results supported the 75

dynamic nature of collective efficacy; as the collective efficacy of the more successful teams 76

increased throughout the race, subsequent performance improved, and vice versa for the less 77

successful teams. 78

How to Measure Collective Efficacy? Resolving the Ambiguity 79

According to the definition of Bandura (1997), efficacy beliefs are future-oriented 80

judgments about capabilities to organize and execute the courses of action. In other words, 81

efficacy measures have to address the skills, properties, or other descriptions of (inter-) 82

personal conditions that assist in successful performance. However, the existing collective 83

efficacy research is characterized by inconsistencies in the manner in which collective 84

efficacy is conceptualized, operationalized, and measured (Shearer, Holmes, & Mellalieu, 85

2009). For instance, current measures of collective efficacy vary with respect to the extent in 86

which they correspond to the original definition of efficacy by Bandura (1997). In line with 87

previous research (Collins & Parker, 2010), we can distinguish two types of measures. 88

The first type evaluates the athletes‟ confidence in their team‟s skills to accomplish the 89

processes that can lead to success (i.e., process-oriented, e.g., “I believe that the players in my 90


team will encourage each other during the game”). Because this type of measure addresses 91

the belief in the team‟s abilities to optimize the process (e.g., items measuring motivational 92

and communication skills that help a team to be successful), it conforms to Bandura‟s original 93

definition of collective efficacy. We will term this measure “collective efficacy” (in the 94

proper process-oriented sense). Collective efficacy thus focuses on athletes‟ confidence in the 95

process of their own team, rather than comparing their own abilities with those of the 96

opposing team. 97

In contrast, the second type of measure focuses on outperforming the opponent and 98

refers to athletes‟ confidence in the abilities of their team to obtain a certain outcome (i.e., 99

outcome-oriented, e.g., “I believe that my team will outplay the opposing team and win this 100

game”). This measure refers to the confidence in the outcome rather than the confidence in 101

the process and focuses on the comparison with the other team, rather than on the own team. 102

Therefore, this measure is not congruent with Bandura‟s original definition of collective 103

efficacy. We will therefore term this outcome-oriented measure “outcome-oriented team 104

confidence”, shortened as “team outcome confidence”. Despite the fact that this outcome-105

oriented team confidence does not measure collective efficacy as originally defined, a number 106

of studies used these measures to allegedly assess collective efficacy (e.g., Chen et al., 2002; 107

Fransen et al., 2012; Spink, 1990; Tasa, Taggar, & Seijts, 2007; Vargas-Tonsing & 108

Bartholomew, 2006). Although previous research (Myers & Feltz, 2007) already 109

recommended against single-item performance measures, typically, the one-item measures 110

used in these studies are outcome-oriented rather than process-oriented, and as such, they 111

measure team outcome confidence rather than collective efficacy (e.g., “What placing do you 112

expect to attain?” or “To what extent do you believe that the team can finish in at least the top 113

10 teams?”). For example, Edmonds and colleagues (2009) attempted to measure the dynamic 114

evolution of collective efficacy in an adventure race by using the one-item measure “How 115


confident are you in the team‟s ability in executing the mountain biking portion of the race in 116

order to secure a top-place finish?” Because this item is more outcome-oriented than process-117

oriented, the authors actually assessed the dynamic variation in team outcome confidence 118

rather than the variation in collective efficacy. 119

Nevertheless, several studies did assess collective efficacy in accordance with the 120

original process-oriented definition of Bandura (1997). An example of a widely used measure 121

of collective efficacy is Short, Sullivan, and Feltz‟s Collective Efficacy Questionnaire for 122

Sport (CEQS; 2005). The CEQS represents collective efficacy as a multidimensional 123

construct based on Bandura‟s (1997) argument that efficacy beliefs include beliefs in the 124

physical tasks but also beliefs in the capability to manage thoughts, actions, emotions, and 125

motivation (Dithurbide & Feltz, 2012, p. 260). The CEQS (2005) comprises a five-factor 126

structure (i.e., five subscales) measured with four items each. These five subscales include: 127

Ability (e.g., “to outplay the opposing team”), Effort (e.g., “to play to its capabilities”), 128

Persistence (e.g., “to persist when obstacles are present”), Preparation (e.g., “to devise a 129

successful strategy”), and Unity (e.g., “to be united”). 130

Given the ambiguity in the current literature concerning the assessment of collective 131

efficacy, the main aim of the present study is to investigate the validity of the measures used 132

to assess collective efficacy. As mentioned above, the one-item measures used to assess 133

collective efficacy often focus on the outcome (i.e., performing better than the opponent), and 134

as such assess outcome-oriented team confidence rather than process-oriented collective 135

efficacy. Consequently, these outcome-oriented one-item measures cannot be used as 136

reference measurement of collective efficacy in team sports. In line with this argument, the 137

validation study by Short and colleagues (2005) revealed a lower correlation between the 138

Ability subscale and the other subscales (.59 - .78) than the correlation among the other 139

subscales (.76 - .94). Looking more closely at the factors and items of the CEQS (Short, et al., 140


2005), it can be inferred that the items of the Ability subscale are outcome-oriented, rather 141

than process-oriented (e.g., “Rate your team‟s confidence, in terms of the upcoming game or 142

competition, that your team has the ability to outplay the opposing team”). Despite the 143

evidence found for the internal consistency of each subscale of the CEQS, the conceptual 144

unity of these different subscales can be questioned. Once clarity is obtained about the 145

reliability of the different collective efficacy measures, the second aim of our study can be 146

realized; the validation of a new and short five-item scale of collective efficacy 147

(Observational Collective Efficacy Scale for Sports; OCESS) that can be used as a starting 148

point for more dynamic measures of collective efficacy. 149

Dynamic Measurements Through Observations: The OCESS 150

While striving toward a more dynamic measurement of collective efficacy, researchers 151

have experienced a practical barrier; in team sports it is not possible to interrupt a player 152

repeatedly during a game to measure his or her collective efficacy beliefs (Myers, Paiement, 153

& Feltz, 2007). Therefore, Edmonds and colleagues (2009) only considered a few time points 154

during a contest. However, in order to advance the knowledge of the dynamic character of 155

collective efficacy, one should strive for more frequent measurements throughout the game. 156

Because working with questionnaires appears to be a major barrier for realizing a continuous 157

measurement of collective efficacy during a contest, observations could provide a viable 158

alternative. 159

A first step toward an observational measure of collective efficacy was taken by 160

Fransen and colleagues (2012). These authors surveyed 33 top-level volleyball coaches on 161

what they believed to be the most important sources of team outcome confidence (i.e., “I 162

believe that my team will win the game”) in their sport. Subsequently, 2365 volleyball 163

coaches and athletes evaluated the extent to which these sources had the power to predict 164

team outcome confidence. The data revealed five sources that were perceived as very 165


important by both coaches and athletes: a) reacting enthusiastically when making a point; b) 166

having leader figures in the team who believe that their team will win this game and express 167

this on the court; c) having both players in the game and on the bench who cheer 168

enthusiastically; d) encouraging each other during the game; and e) communicating tactically 169

during the game. All these behaviors are clearly process-oriented. Having confidence that the 170

own team has the qualities to succeed in these five behaviors could therefore represent 171

process-oriented collective efficacy. 172

In the present study we develop a new scale based on these five sources, named the 173

Observational Collective Efficacy Scale for Sports (OCESS). The aim of the present study is 174

to assess whether this short scale constitutes a valid measure of process-oriented collective 175

efficacy in different team sports. If it does, the 5-item OCESS would offer a valid alternative 176

to the 20-item CEQS for assessing collective efficacy in sport whenever time available for 177

administering long questionnaires is limited. Furthermore, because all five items represent 178

observable behaviors, the OCESS would allow future assessment of the evolution of players‟ 179

collective efficacy beliefs throughout a contest by observations rather than questionnaires. 180

Such a measure could highlight the dynamic nature of collective efficacy during a game and 181

provide more insight into how to attain and maintain high collective efficacy. 182

Hypotheses 183

Given the ambiguity in the existing literature concerning the assessment of collective 184

efficacy, the main purpose of the present study is to investigate the validity of the measures 185

currently used to assess collective efficacy in sports teams. In line with our conceptual 186

reasoning above, we hypothesize that the Ability subscale assesses outcome-oriented team 187

confidence (analogous to the outcome-oriented one-item measures), rather than process-188

oriented collective efficacy. By contrast, we expect the other four subscales of the CEQS to 189


form a valid and reliable reference measurement of process-oriented collective efficacy as 190

defined by Bandura (1997). 191

Once a reliable reference measurement of collective efficacy is obtained, a second aim 192

of our study can be realized: the validation of our newly developed five-item scale of 193

collective efficacy (Observational Collective Efficacy Scale for Sports; OCESS) within 194

different team sports. Two hypotheses can be formulated with regard to this aim. First, we 195

hypothesize that the OCESS and the CEQS (subscales 2-5) are strongly correlated (i.e., r > 196

.70), attesting that the OCESS measures process-oriented collective efficacy instead of 197

outcome-oriented team confidence. Second, the convergent and divergent validity of the 198

OCESS is examined by comparing the influence of demographic characteristics respectively 199

with the first subscale and the last four subscales of the CEQS. If supported, this OCESS, 200

which includes only observable behaviors, offers a starting point for the design of a 201

continuous measure of players‟ collective efficacy beliefs during the game through 202

observation instead of through the use of traditional questionnaires. 203

Method 204

Procedure 205

The database of the Flemish Trainer School (i.e., organization responsible for sport-206

specific schooling of coaches in Flanders) was used to invite 5,535 qualified coaches out of 207

nine different team sports to participate in our study. These coaches were asked to complete a 208

web-based questionnaire and to motivate their players to complete the player-specific version 209

of the questionnaire. In order to assure variability within our sample, we also contacted non-210

qualified coaches and their teams through the different Flemish sport federations. The coaches 211

and players who did not respond were sent a reminder two weeks later. Informed consent was 212

obtained from all participants. No rewards were given for participation in our study and all 213

participants were guaranteed full confidentiality. 214


Participants 215

In total, 4,451 participants (3,193 players and 1,258 coaches) completed our 216

questionnaire. This corresponds to an approximate response rate of 27%. These participants 217

played or coached in 2,366 different teams. More detailed information on the participants can 218

be found in Table 1. The sample included participants from nine team sports in Flanders; 219

basketball, handball, hockey, ice hockey, netball, rugby, soccer, volleyball, and water polo. 220

Table 2 contains the descriptive characteristics for the respondents of each of the nine team 221

sports. Data from this sample have been used in another research study (Fransen, 222

Vanbeselaere, De Cuyper, Vande Broek, & Boen, 2014), but examined different variables and 223

research questions. 224

Measures 225

Collective efficacy. Two measures of collective efficacy were included in our 226

questionnaire. First, the Collective Efficacy Questionnaire for Sports (CEQS; Short, et al., 227

2005), including five subscales, each consisting of four items. In line with the suggestions of 228

Myers and Feltz (2007), each of the items begins with the stem: “Rate your confidence, in 229

terms of the upcoming game or competition, that your team has the ability to…” Participants 230

assessed the items on a 7-point scale anchored by 1 (not at all confident) and 7 (extremely 231

confident). 232

The second collective efficacy measure included in our study was our newly 233

developed five-item Observational Collective Efficacy Scale for Sports (OCESS), including 234

the most important sources of team outcome confidence (Fransen, et al., 2012). It is important 235

to note that, although the items of the OCESS are intended to be used as an observational 236

measurement instrument in the future, in the current study, the scale is still in a self-evaluative 237

questionnaire form. The items included in the OCESS are “react enthusiastically when 238

making a point,” “have leader figures in the team who believe that we will win this game and 239


express this on the court,” “have both players in the game and on the bench who cheer 240

enthusiastically,” “encourage each other during the game,” and “communicate a lot tactically 241

during the game.” In analogy with the CEQS, each of the items was assessed on a 7-point 242

scale ranging from 1 (not at all confident) to 7 (extremely confident) and each item began with 243

the stem: “Rate your confidence, in terms of the upcoming game or competition, that your 244

team has the ability to…” 245

Team outcome confidence. Outcome-oriented team confidence was measured using 246

five one-item measures that assess the confidence that the team will win the game, lose the 247

game, or realize its goals. These items are a general representation of the measures mainly 248

used in previous research studies (Myers & Feltz, 2007, for a review). To determine the 249

difference between an individual stem (i.e., “I believe that our team…”) and the team-focused 250

stem (i.e., “Our team believes that we…”), we included items with both stems for the 251

confidence in winning or losing the upcoming game. 252

Other measures. Besides several background characteristics (e.g., sex, age, years of 253

experience), we assessed some performance related measures as well, such as position of the 254

team in the ranking of the ongoing season and the score and quality of the play during the last 255

game. 256

Results 257

In order to validate our new OCESS scale as a measure of collective efficacy in sports 258

teams, we first investigated the validity of the measures currently used to assess collective 259

efficacy for the Flemish context. 260

Investigation of the Validity of the Flemish Version of the Collective Efficacy 261

Questionnaire for Sports (CEQS) 262

Factor analyses. A Confirmatory Factor Analysis (CFA) conducted on the 20-item 263

CEQS questionnaire, including the five subscales, for all 4,451 players and coaches, revealed 264


an inadequate fit with the data (χ² = 5620; df = 165; p < .001; GFI = .87; AGFI = .84; RMSEA 265

= .09). We therefore conducted an Exploratory Factor Analysis on the whole sample (4,451 266

players and coaches within all sports) to identify the structure underneath the 20 items of the 267

CEQS scale. It has been established that the scree plot is a reliable criterion for component 268

selection with samples of more than two hundred participants (Stevens, 2002). The scree plot 269

suggested that two independent factors should be extracted which explained 61% of variance. 270

An item was retained to construct a factor when it had a minimum loading of .40, without 271

having a cross loading higher than .40 on another factor. This resulted in the deletion of three 272

items from different subscales; the items “Be ready” and “Devise a successful strategy” were 273

deleted from the subscale Preparation, the item “Perform under pressure” was deleted from 274

the subscale “Persistence”. The first component, accounting for 52% of the variance in 275

participants‟ responses, consisted of 13 items from the subscales of Effort, Persistence, 276

Preparation, and Unity. The second component included the four items of the CEQS subscale 277

of Ability. 278

Intercorrelations between the subscales of the CEQS. In order to provide a better 279

insight into the underlying structure of the five subscales of the original CEQS, Table 3 280

presents the correlation matrix of all subscales of the CEQS scale. Cronbach‟s α coefficients 281

are provided in parentheses on the diagonal as estimates of internal consistency. 282

The internal consistency of all five subscales was high (all Cronbach‟s α‟s > .83). As 283

can be seen in Table 3, subscales 2, 3, 4, and 5 are strongly correlated (all r > .69). However, 284

the Ability subscale is only moderately correlated (i.e., r < .60) with the other subscales. This 285

confirms the previous EFA that this subscale measures something different than the other 286

subscales. 287

The relation between CEQS and outcome-oriented team confidence. To 288

investigate the internal validity of the different subscales of the CEQS we explore the 289


relationship with five one-item measures of outcome-oriented team confidence. Table 4 290

presents all correlations between these five one-item measures and the five subscales of the 291

CEQS (Short, et al., 2005). 292

The outcome-oriented beliefs (i.e., winning/losing the game) correlate strongly with 293

the Ability subscale. Also, the item assessing the belief in obtaining a goal correlates more 294

strongly with the Ability subscale than with the other four subscales. The subscales Effort, 295

Persistence, Preparation, and Unity correlate only moderately with outcome-oriented team 296

confidence (all r < .49). The internal consistency of this newly constructed scale (subscales 2-297

5 of the CEQS) is very high (Cronbach‟s α = .95). Additional analyses revealed high 298

correlation between the items: “I believe that our team will win the game” and “Our team 299

believes that we will win the game” (r = .80; p < 0.01). 300

The Observational Collective Efficacy Scale for Sports (OCESS) 301

The findings above make clear that the subscales Effort, Persistence, Preparation, and 302

Unity of the CEQS form a reliable measure of process-oriented collective efficacy. This 303

brings us to the second purpose of our study, namely to determine whether our newly 304

developed five-item OCESS can be considered as an adequate measure for process-oriented 305

collective efficacy. The Cronbach‟s α of the 5-item OCESS is .85, indicating a high internal 306

consistency. 307

Correlation with CEQS. Table 5 shows the correlations between the OCESS and the 308

CEQS, including correlations with the full scale as well as correlations with the different 309

subscales. In addition, the correlation with the process-oriented part of the CEQS (subscales 310

2-5) is reported. The results reveal high correlations between the OCESS and CEQS subscales 311

2, 3, 4, and 5, which together represent the process-oriented part of the CEQS (r = .79). In 312

contrast, only a moderate correlation with the CEQS Ability subscale emerged. 313


Relation with demographic variables. In order to further test the validity of the 314

OCESS as measure of collective efficacy, we explored both convergent and discriminant 315

validity by comparing the influence of demographic variables on different scales. With regard 316

to the convergent validity, we tested whether the OCESS and the process-oriented part of the 317

CEQS (subscales 2-5) are similarly related with the demographic variables. To examine the 318

discriminant validity, we tested whether the OCESS and the first subscale of the CEQS (as 319

measure of the outcome-oriented team confidence) are related with the predictors in a 320

different way. 321

We conducted three regression analyses with the different demographic variables as 322

predictors (see Table 6). The Ability subscale of the CEQS (presumably a measure of team 323

outcome confidence), the process-oriented part of the CEQS (subscales 2-5), and the newly 324

developed OCESS served as criterion variables. Because the large sample size (N = 4450) 325

goes along with an extremely high statistical power, we will consider only the significant 326

relationships with a β-value above .20 (explaining at least 4% of the variance). The regression 327

analyses in Table 6 reveal that the different demographic characteristics have a very similar 328

relation with the two criteria that we consider as measures of collective efficacy (i.e., 329

subscales 2-5 of the CEQS and the OCESS). Both the place in ranking of the own team and 330

the playing level of the own team in the game of last weekend are significantly, and in the 331

same direction, related with the two collective efficacy scales, which supports the convergent 332

validity of our OCESS scale. By contrast, two different demographic variables, namely the 333

place in the ranking of the next game‟s opponent and the score of the first game against that 334

opponent, were significantly related to outcome-oriented team confidence. This differential 335

impact of demographic variables supports the discriminant validity of the OCESS scale. 336

337

338


Discussion 339

The results of the present study question the internal validity of the measures currently 340

used to assess collective efficacy. Two types of measures could be distinguished: process-341

oriented collective efficacy (i.e., the confidence in the team‟s skills to accomplish the 342

processes that could lead to successes) and outcome-oriented team confidence (i.e., the 343

confidence in the team‟s ability to obtain a goal or win a game). Furthermore, our findings 344

provide support for our contention that the developed five-item OCESS can be used as a valid 345

measure of process-oriented collective efficacy. 346

First, the results of this study demonstrated that the internal consistency of each of the 347

five subscales of the Collective Efficacy Questionnaire for Sports (Short, et al., 2005), as well 348

as the internal consistency of the full scale, was high. On the other hand, the originally 349

proposed five-factor structure showed only a moderate fit to the data. The Ability subscale 350

emerged as a separate factor with relatively lower correlations with the other subscales, and 351

with different relations with the demographic variables. This Ability subscale was found to 352

assess outcome-oriented team confidence, rather than process-related collective efficacy, 353

given its high correlations with the outcome-oriented one-item measures. The combined 354

subscales Effort, Persistence, Preparation, and Unity seem to constitute a measure for process-355

related collective efficacy. Both findings are in line with our hypothesis. 356

Second, the present findings suggest that the OCESS is a valid measure of process-357

oriented collective efficacy in different team sports. First, the OCESS scale has a high internal 358

consistency. Second, high correlations have been established with the four subscales of the 359

CEQS that assess process-oriented collective efficacy (r > .68). In contrast, only a moderate 360

correlation emerged with the Ability subscale. This indicates that the OCESS is a measure of 361

process-oriented collective efficacy rather than a measure of outcome-oriented team 362

confidence. The convergent validity of the OCESS was further supported by the similar 363


relations between demographic characteristics and both the OCESS scale and the process-364

oriented part of the CEQS. In contrast, these demographic characteristics had different 365

relations with the Ability subscale, supporting the discriminant validity, and providing further 366

evidence that the Ability subscale of the CEQS does not measure process-oriented collective 367

efficacy beliefs that are congruent with Bandura‟s (1997) definition of the construct. 368

In addition, in this original definition, Bandura (1997) referred to collective efficacy as 369

“a group‟s shared belief”. Nevertheless, previous research argued that the best way to capture 370

efficacy beliefs in questionnaires is by assessing the individual‟s perception of the team‟s 371

capabilities (Bandura, 1997; Myers & Feltz, 2007; Shearer, Holmes, & Mellalieu, 2009). It 372

should be noted that the OCESS contains items that express interaction or interpersonal 373

behavior (e.g., communicating tactically, encouraging each other). These behaviors can be 374

interpreted as “shared” behavior, and therefore align more closely with the original definition 375

of Bandura (1997). 376

Because all the items in the OCESS refer to behaviors that can be observed, this scale 377

offers a starting point for the development of a continuous observational instrument of 378

collective efficacy during a competitive game. Because this new measure of collective 379

efficacy can be completed by observers, it has the potential to overcome the limitations of 380

traditional questionnaires that have to be completed by the players themselves. Moreover, 381

such observations allow assessing the dynamical changes of collective efficacy (e.g., in 382

critical periods during a game). 383

Our study includes strengths and limitations, so the results should be interpreted 384

accordingly. A particular strength of the study is the large sample size of both coaches and 385

athletes, as well as the diversity of sport and competition level. Having such a large and 386

diverse sample increases the applicability of the results to various sport settings. In addition, 387


the five-item OCESS offers a valid alternative to one-item measures for assessing collective 388

efficacy in sport whenever time available for administering long questionnaires is limited. 389

A potential limitation associated with our study is the use of an online survey to gather 390

the data, which resulted in participation of individual players and coaches rather than 391

complete teams. Because the 4,451 participants were active in 2,366 different teams, it was 392

not possible to establish whether these collective efficacy beliefs are shared within the team. 393

Collective efficacy is a group-level construct that is typically measured at the individual level 394

and then, when appropriate, aggregated to the group or team level for subsequent analysis. 395

This study only measured collective efficacy beliefs at the individual level of analysis. Further 396

research is required to explore whether a similar pattern will be obtained at the group-level of 397

analysis. 398

A second limitation regards to the design of our study. Given our cross-sectional study 399

design, we are not able to give evidence for the amount of stability or instability of the 400

OCESS over time. Because the OCESS (in an observational form) should be able to capture 401

changes in collective efficacy (e.g., during a game or between subsequent games), the 402

measurement has to be sensitive for variations. On the other hand, given the stability of 403

external and internal circumstances, we expect high test-retest-reliability. More clarity should 404

be obtained with further studies. 405

Another suggestion for future research refers to the validation of the OCESS as 406

observational measure of collective efficacy. The present manuscript provides the first 407

necessary step in this validation process by demonstrating that the self-reported efficacy 408

behaviors (i.e., the OCESS) are highly correlated with collective efficacy, as measured by the 409

process-oriented part of the CEQS. Future work is required to complete the final step in this 410

validation process, namely to establish a high correlation between the self-reported efficacy 411

behaviors and the observer-reported efficacy behaviors in a real game setting (both assessed 412


by the OCESS). To obtain a high inter-observer reliability, it will be essential to define and 413

standardize the observation of the five behaviors for each specific sport, as well as to train the 414

observers in this behavioral assessment. 415

The findings of the present study contribute both to theoretical knowledge and to 416

coaching practice. First, the results provide clear insight into the conceptual distinction 417

between process-oriented collective efficacy and outcome-oriented team outcome confidence. 418

Hopefully, these findings result in more conceptual clarity in future collective efficacy 419

research. Furthermore, these findings have the potential to provide the basis for the 420

development of a dynamic collective efficacy measurement based on observations guided by 421

the OCESS. Such a measure could provide a better insight in the dynamic nature of collective 422

efficacy during a game and its relation with performance. 423

Second, this continuous measure would constitute an added value for the coaching 424

practice by providing coaches with more insights into how to attain and maintain high 425

collective efficacy standards within their teams. In addition to technical and tactical scouting, 426

this mental scouting of players can become an essential tool to make important decisions in 427

the course of a game. 428


References 429

Bandura, A. (1997). Self-efficacy: The exercise of control. New York: Freeman. 430

Chen, G., Webber, S. S., Bliese, P. D., Mathieu, J. E., Payne, S. C., Born, D. H., & Zaccaro, 431

S. J. (2002). Simultaneous examination of the antecedents and consequences of 432

efficacy beliefs at multiple levels of analysis. Human Performance, 15(4), 381-409. 433

Collins, C. G., & Parker, S. K. (2010). Team capability beliefs over time: Distinguishing 434

between team potency, team outcome efficacy, and team process efficacy. Journal of 435

Occupational and Organizational Psychology, 83(4), 1003-1023. doi: 436

10.1348/096317909x484271 437

Dithurbide, L., & Feltz, D. L. (2012). Self-efficacy and collective efficacy. In G. Tenenbaum, 438

R. C. Eklund & A. Kamata (Eds.), Measurement in sport and exercise psychology (pp. 439

251-263). Champaign, IL: Human Kinetics. 440

Dithurbide, L., Sullivan, P., & Chow, G. (2009). Examining the influence of team-referent 441

causal attributions and team performance on collective efficacy: A multilevel analysis. 442

Small Group Research, 40(5), 491-507. doi: 10.1177/1046496409340328 443

Edmonds, W. A., Tenenbaum, G., Kamata, A., & Johnson, M. B. (2009). The role of 444

collective efficacy in adventure racing teams. Small Group Research, 40(2), 163-180. 445

doi: 10.1177/1046496408328489 446

Fransen, K., Vanbeselaere, N., Exadaktylos, V., Vande Broek, G., De Cuyper, B., Berckmans, 447

D., . . . Boen, F. (2012). "Yes, we can!": Perceptions of collective efficacy sources in 448

volleyball. Journal of Sports Sciences, 30(7), 641-649. doi: 449

10.1080/02640414.2011.653579 450

Fransen, K., Vanbeselaere, N., De Cuyper, B., Vande Broek, G., & Boen, F. (2014). The myth 451

of the team captain as principal leader: Extending the athlete leadership classification 452


within sport teams. Journal of Sports Sciences, 1-9. doi: 453

10.1080/02640414.2014.891291 454

Greenlees, I. A., Graydon, J. K., & Maynard, I. W. (1999). The impact of collective efficacy 455

beliefs on effort and persistence in a group task. Journal of Sports Sciences, 17(2), 456

151-158. 457

Hodges, L., & Carron, A. V. (1992). Collective efficacy and group performance. International 458

Journal of Sport Psychology, 23(1), 48-59. 459

Keshtan, M. H., Ramzaninezhad, R., Kordshooli, S. S., & Panahi, P. M. (2010). The 460

relationship between collective efficacy and coaching behaviors in professional 461

volleyball league of Iran clubs. World Journal of Sport Sciences, 3(1), 1-6. 462

Mangan, A. (2013, April 5). Have Arsenal found confidence at the right time? Retrieved 463

from http://espnfc.com 464

Myers, N. D., & Feltz, D. L. (2007). From self-efficacy to collective efficacy in sport: 465

Transitional methodological issues. In G. Tenenbaum & R. C. Eklund (Eds.), 466

Handbook of sport psychology (3rd ed., pp. 799-819). Hoboken, NJ, US: John Wiley 467

& Sons Inc. 468

Myers, N. D., Feltz, D. L., & Short, S. E. (2004). Collective efficacy and team performance: 469

A longitudinal study of collegiate football teams. Group Dynamics: Theory, Research, 470

and Practice, 8(2), 126-138. doi: 10.1037/1089-2699.8.2.126 471

Myers, N. D., Paiement, C. A., & Feltz, D. L. (2007). Regression team performance on 472

collective efficacy: Considerations of temporal proximity and concordance. 473

Measurement in Physical Education and Exercise Science, 11(1), 1-24. 474

Myers, N. D., Payment, C. A., & Feltz, D. L. (2004). Reciprocal relationships between 475

collective efficacy and team performance in women's ice hockey. Group Dynamics: 476

Theory, Research, and Practice, 8(3), 182-195. doi: 10.1037/1089-2699.8.3.182 477


Shearer, D. A., Holmes, P., & Mellalieu, S. D. (2009). Collective efficacy in sport: the future 478

from a social neuroscience perspective. International Review of Sport and Exercise 479

Psychology, 2(1), 38-53. doi: 10.1080/17509840802695816 480

Short, S. E., Sullivan, P., & Feltz, D. (2005). Development and preliminary validation of the 481

collective efficacy questionnaire for sports. Measurement in Physical Education and 482

Exercise Science, 9(3), 181-202. 483

Silver, W. S., & Bufanio, K. M. (1996). The impact of group efficacy and group goals on 484

group task performance. Small Group Research, 27(3), 347-359. 485

Spink, K. S. (1990). Group cohesion and collective efficacy of volleyball teams. Journal of 486

Sport & Exercise Psychology, 12(3), 301-311. 487

Stevens, J. (2002). Applied multivariate statistics for the social sciences (4th ed.). Hillsdale, 488

NJ: Erlbaum. 489

Tasa, K., Taggar, S., & Seijts, G. H. (2007). The development of collective efficacy in teams: 490

A multilevel and longitudinal perspective. Journal of Applied Psychology, 92(1), 17-491

27. doi: 10.1037/0021-9010.92.1.17 492

Vargas-Tonsing, T. M., & Bartholomew, J. B. (2006). An exploratory study of the effects of 493

pregame speeches on team efficacy beliefs. Journal of Applied Social Psychology, 494

36(4), 918-933. 495


Table 1 496

Sample characteristics 497

Participants MAge

(years)

MExperience

(years)

Team gender Level

Coaches 1,258 (28%) 41.94 13.97 905 ♂ (72%)

353 ♀ (28%)

90 E

268 N

613 P

102 RG

22 RC

163 Y

(7%)

(21%)

(49%)

(8%)

(2%)

(13%)

Players 3,193 (72%) 23.92 14.21 1,915 ♂ (60%)

1,278 ♀ (40%)

177 E

836 N

1,733 P

209 RG

122 RC

116 Y

(6%)

(26%)

(54%)

(7%)

(4%)

(4%)

Total sample 4,451 29.01 14.14 2,820 ♂ (63%)

1,631 ♀ (37%)

267 E

1,104 N

2,346 P

311 RG

144 RC

279 Y

(6%)

(25%)

(53%)

(7%)

(3%)

(6%)

Note. ♂ = male; ♀ = female; E = elite level; N = national level; P = provincial level; RG = 498

regional level; RC = recreational level; Y = youth teams. 499


Table 2 500

Sport specific sample characteristics 501

Participants Mage

(years)

MExperience

(years)

Male team (♂) /

Female team (♀)

Function

Players (P) /

Coaches (C)

Basketball 1,959 (44%) 27.40 14.67 1,332 ♂ (68%)

627 ♀ (32%)

1,551 P (79%)

408 C (21%)

Volleyball 1,287 (29%) 29.77 14.35 521 ♂ (41%)

766 ♀ (59%)

919 P (71%)

368 C (29%)

Soccer 589 (13%) 33.88 13.05 541 ♂ (92%)

48 ♀ (8%)

249 P (42%)

340 C (58%)

Hockey 127 (3%) 27.39 13.65 68 ♂ (53%)

59 ♀ (47%)

110 P (87%)

17 C (13%)

Netball 118 (3%) 27.53 15.27 64 ♂ (54%)

54 ♀ (46%)

85 P (72%)

33 C (28%)

Handball 116 (3%) 29.64 13.67 80 ♂ (69%)

36 ♀ (31%)

76 P (65%)

40 C (35%)

Water polo 99 (2%) 26.93 13.40 84 ♂ (85%)

15 ♀ (15%)

84 P (85%)

15 C (15%)

Rugby 84 (2%) 28.10 7.59 67 ♂ (80%)

17 ♀ (20%)

60 P (71%)

24 C (29%)

Ice hockey 72 (2%) 27.76 13.37 63 ♂ (87%)

9 ♀ (13%)

59 P (82%)

13 C (18%)

Total sample 4,451 29.01 14.14 2,820 ♂ (63%)

1,631 ♀ (37%)

3,193 P (72%)

1,258 C (28%)

502


Table 3 503

Intercorrelations between different subscales of the CEQS (Short et al., 2005). The 504

Cronbach’s α coefficient of each subscale can be found on the diagonal in parentheses. 505

Subscale 1

Ability

Subscale 2

Effort

Subscale 3

Persistence

Subscale 4

Preparation

Subscale 5

Unity

Subscale 1 Ability (.93)

Subscale 2 Effort .51**

(.83)

Subscale 3 Persistence .56**

.79**

(.83)

Subscale 4 Preparation .59**

.75**

.69**

(.84)

Subscale 5 Unity .52**

.80**

.79**

.73**

(.84)

**p < .01 506


Table 4 507

Correlations between the subscales of the CEQS and five one-item measures of outcome-508

oriented team confidence 509

Subscale 1

Ability

Subscale 2

Effort

Subscale 3

Persistence

Subscale 4

Preparation

Subscale 5

Unity

I believe that our team will

win the upcoming game

.77**

.37**

.40**

.44**

.38**


lose the upcoming game

-.73**

-.34**

-.37**

-.41**

-,35**


obtain its goal in the

upcoming game

.59**

.47**

.48**

.49**

.49**

Our team believes that we

will win the upcoming game

.75**

.40**

.44**

.48**

.41**

Our team believes that we

will lose the upcoming game

-.69**

-.35**

-.39**

-.43**

-.36**

** p < .01 510


Table 5 511

The correlations between the five-item OCESS (both full scale and individual items) and the 512

CEQS (Short et al., 2005) 513

Full

CEQS

S1

Ability

S2

Effort

S3

Persistence

S4

Preparation

S5

Unity

S2-5

Full OCESS .78**

.51**

.75**

.68**

.68**

.75**

.79**

1. React enthusiastically

when making a point

.51**

.27**

.55**

.45**

.46**

.50**

.54**

2. Have leader figures in

the team who believe that

we will win this game and

express this on the court

.62**

.51**

.56**

.52**

.52**

.55**

.59**

3. Have both players in the

game and on the bench

who cheer enthusiastically

.61**

.37**

.62**

.56**

.51**

.60**

.63**

4. Encourage each other

during the game

.64**

.36**

.64**

.57**

.53**

.65**

.66**

5. Communicate a lot

tactically during the game

.66**

.45**

.57**

.57**

.64**

.63**

.67**

**p < .01 514


Table 6 515

Regression analyses with background characteristics as predictors and CEQS and OCESS as 516

dependent variables. The significant beta values are marked in bold. 517

Predictors

CEQS

Subscale 1

Team outcome

confidence

R² = .391

CEQS

Subscale 2-5

Collective efficacy

R² = .180

OCESS

Collective efficacy

R² = .130

β β β

Player/Coach .06* .13

*** .04

Sex .03 -.06 -.10**

Male/Female team .01 .04 .05

Age -.04 .02 .03

Years of experience .01 .01 -.02

Team level -.04**

-.10***

-.05**

Team tenure .02 .05**

.06**

Place in ranking of own team -.18***

-.25***

-.25***

Place in ranking of opponent -.33***

-.07**

-.02

Score of first game against

same opponent

.20***

-.00 .00

Score of game last weekend .03 -.01 -.00

Ranking opponent of game last

weekend

.04* -.02 .00

Playing level own team game

last weekend .10

*** .23

*** .20

***

*p < .05

**p < .01

***p < .001 518

international journal of sport psychology, 45 · pdf file1 international journal of sport...

Documents